In a simple linear regression $Y_i = \beta_0 + \beta_1 X_i + \epsilon_i$, suppose the error term variance is heteroscedastic such that $Var(\epsilon_i|X_i) = \sigma^2 X_i^2$. If the Ordinary Least Squares (OLS) estimator $\hat{\beta}_1$ is used, which statement is true regarding its properties?